Statistics and Data Analysis
Code: 44079 ECTS Credits: 9| Degree | Type | Year | Semester |
|---|---|---|---|
| 4313861 High Energy Physics, Astrophysics and Cosmology | OB | 0 | 1 |
Contact
- Name:
- Francisco Javier Rico Castro
- Email:
- franciscojavier.rico@uab.cat
Use of Languages
- Principal working language:
- english (eng)
Teachers
- Ramon Miquel Pascual
- Abelardo Moralejo Olaizola
- Jorge Carretero Palacios
- Pau Tallada Crespí
- Martin Borstad Eriksen
- Francesc d'Assis Torradeflot Curero
Prerequisites
For the Python Bootcamp (part 2), it is highly needed to bring a personal laptop with a running installation of Python 3.9.
For that, install Python 3.9 with the Anaconda installer. In this way, your Python distribution will contain all the associated packages needed for this course.
Follow these steps:
-
Download Anaconda installer for Python 3.9 here https://www.anaconda.com/download/
-
Follow the installation instructions - both GUI or terminal versions work fine. If prompted, select the option to add the new anaconda directory to your path.
The use of Linux or Mac is highly recommended.
Objectives and Contextualisation
In this course we will learn how to distill scientific knowledge from experimental data, a process that relies on statistical methods. We will learn the basics concepts of Probability and Statistics (in their Frequentist and Bayesian frameworks). In addition, we will study and practice several particular statistical methods and data analysis techniques usually used in the fields of High Energy Physics, Astrophysics and Cosmology. To that aim, we will learn and practice the use of modern statistics and analysis software tools.
Competences
- Solve problems in new or little-known situations within broader (or multidisciplinary) contexts related to the field of study.
- Use mathematics to describe the physical world, select the appropriate equations, construct adequate models, interpret mathematical results and make critical comparisons with experimentation and observation.
- Use the adequate software, programming languages and computer packages to research problems related to high energy physics, astrophysics and cosmology.
- Work in a group and take on responsibility, interacting professionally and constructively with other people with complete respect for their rights.
Learning Outcomes
- Apply data analysis techniques to problems in the areas of particle physics, astrophysics and cosmology, as well as other close but different areas.
- Learn how statistical analysis software works.
- Use Monte Carlo techniques to model real problems of physics.
- Work in small groups to solve problems of data analysis.
Content
Part 1: Basic concepts on probability, statistics and Monte Carlo techniques
Part 2: Python for Statistics and Data Analysis
Part 3: Parameter estimation, Hypothesis test and Unfolding
Part 4: Bayesian Statistics
Methodology
- Theory lectures including practical examples in the fields of High Energy Physics, Astrophysics and Cosmology
- Homework exercises to be solved by students alone or in small groups
- Discussion of problems during classes and tutorials
- Hands-on sessions on software tools for statistics and data analysis (in Python programming language)
- Explanation and discussion of sample code/algorithms in Python programming languages during classes and tutorials
Annotation: Within the schedule set by the centre or degree programme, 15 minutes of one class will be reserved for students to evaluate their lecturers and their courses or modules through questionnaires.
Activities
| Title | Hours | ECTS | Learning Outcomes |
|---|---|---|---|
| Type: Directed | |||
| Lectures | 56 | 2.24 | 1, 2, 3 |
| Study of theory and practical examples | 40 | 1.6 | 1, 2, 4, 3 |
| Type: Autonomous | |||
| Discussion, workgroups, problem solving | 34 | 1.36 | 1, 2, 4, 3 |
Assessment
The evaluation will take into account:
- Attendance and active participation to the lectures
- Resolution, for each of the course parts, of specific take-home exercises
- Resolution of a final, synthesis take-home exam
For those students not passing the course after the regular evaluation procedure, there will be a recuperation evaluation round consisting also on specific take-home exercises for the different course parts, plus a final, synthesis exam. There will be no threshold mark to be eligible for the recuperation evaluation round, other than the general requirement of having been evaluated at least for a 66% of the total qualification activities in the first round.
Assessment Activities
| Title | Weighting | Hours | ECTS | Learning Outcomes |
|---|---|---|---|---|
| Attendance and active participation to the lectures | 5% | 0 | 0 | 1, 2, 3 |
| Resolution of a final, synthesis exam | 50% | 50 | 2 | 1, 2, 4, 3 |
| Resolution of class exercises | 45% | 45 | 1.8 | 1, 2, 4, 3 |
Bibliography
- G. Cowan; "Statistical Data Analysis", 1998, Oxford University Press
- P. A. Zyla et al. (Particle Data Group); "Review of Partcle Physcis (2021)", Prog. Theor. Exp. Phys. 2020, 083C01 (2020) and 2021 update
- F. James; "Statistical Methods in Experimental Physics", 2nd Edition, 2006, World Scientific
- L. Lyons, "Statistics for Particle and Nuclear Physicists", 1986, Cambridge University Press
- B. P. Roe, "Probability and Statistics in Experimental Physics", 1992, Springer
- A. G. Frodesen, et al., "Probability and statistics in particle physics", 1979, Columbia University Press D. Sivia and J. Skilling, "Data Analysis, A Bayesian Tutorial", 2nd ed., 2006, Oxford University Press A. Gelman, "Bayesian Data Analysis", 1995, CRC Press
- R. J. Barlow, "Statistics", 1989, J. Wiley
- W.T. Press et al., "Numerical Recipes: The Art of Scientific Computing", Cambridge University Press. E.T. Jaynes, "Probability Theory: The Logic of Science", Cambridge University Press.
- A. Stuart et al., "Kendall's Advanced Theory of Statistics", Vol 2A. Wiley.
- F. James, "Monte Carlo Theory and Practice", Rep. Prog. Phys. 43 (1980) 73.
Software
We will introduce and make use of the Python programming language (see the "Prerequisists" section for installation details).
In particular, we will study and use the following Python libraries: numpy, pandas, matplotlib, scipy and scikit learn.